How a new class of startups are working to solve the grid storage puzzle

Here’s the problem: Solar panels and wind turbines are cheap, clean, reliable sources of electricity, right up until they’re not. The sun sets; the wind flags. They can’t power an electricity grid alone.

Coal and natural-gas plants can fill in the gaps today. But as climate regulations shutter more of these carbon-spewing sources, there will eventually be days or even weeks each year when renewables won’t be enough to keep the lights on. Something else will need to step in.

Form Energy is convinced that that something could be a battery. But it’d have to be a battery unlike any the world has seen.

To be as cheap, reliable, and flexible as natural gas, such a battery system would have to cost less than $10 per kilowatt-hour. Today’s best grid batteries, large lithium-ion systems, cost hundreds of dollars per kilowatt-hour (precise estimates vary). It could take decades even for that price to drop below $100.

It’s a huge leap. But Form’s founders think they could hit that target by developing big batteries that rely on extremely cheap, energy-dense materials. “We think we can get there,” says MIT professor Yet-Ming Chiang, cofounder and chief scientist at Form. “We think we can match technology to those requirements.”

A low-cost, long-lasting form of energy storage that could be built anywhere would be about the closest thing to a silver bullet for cleaning up the power sector. It would make the most of the sharply declining costs of solar and wind, without many of the environmental, safety, or aesthetic problems raised by other ways of balancing out fluctuating renewables.

Form, based in Somerville, Massachusetts, seized the attention of the battery world when it was created in 2017. Chiang is one of the world’s top battery scientists. He’s published hundreds of scientific papers, holds more than 80 patents, and has cofounded six startups. Several have earned valuations of more than $1 billion, including A123 Systems, which makes lithium-ion batteries for electric vehicles.

Form’s CEO, Mateo Jaramillo, previously assembled and led a business unit of Tesla that sells battery systems for homes and is now building some of the largest grid battery projects around the world. To date, Form has raised around $50 million from Bill Gates’s Breakthrough Energy Ventures, Italian energy giant Eni, and others.

A wave of earlier grid storage companies failed (See “Why bad things happen to clean-energy startups”). Form is just one of several that have recently raised funds to take a fresh crack at the problem.

The main storage need on the grid today is known as “intraday storage.” It provides quick bursts of electricity for a few hours to smooth out mismatches between generation and demand throughout the day and at least into the early evening.

A growing amount of that storage comes from lithium-ion batteries, which also power phones, laptops, and electric cars and are steadily getting cheaper and more powerful. The amount of grid energy storage installed globally rose almost 150% last year to six gigawatt-hours, according to research firm Wood Mackenzie. That's nearly double the average rate during the preceding five years, and lithium-ion systems accounted for most of the increase.

Tesla, for instance, plans to build hundreds of its new three-megawatt-hour Megapack battery systems in Moss Landing, California. The project, which includes other energy storage developers as well, would replace a trio of decades-old gas plants at the site run by Calpine, a large American power company.

Meanwhile, a growing number of renewables developers, like Recurrent Energy and First Solar, are proposing giant solar farms coupled with huge battery storage systems, enabling the plants to continue delivering electricity for hours after sunset.

But the sun and wind don’t just fade for hours; sometimes they dip for days or weeks. If we want to shift mainly to renewables, we’re going to need a lot more storage that can last a lot longer.

With today’s battery technology, the costs would skyrocket, says Jesse Jenkins, an assistant professor at Princeton who researches energy systems. It would require banks upon banks of lithium-ion batteries, many of which might be used only a few times a year. We’d also need to build more solar and wind farms to generate enough surplus electricity to charge them. (See “The $2.5 trillion reason we can’t rely on batteries to clean up the grid.”)

The economics crumble in this scenario. “If these assets are supposed to lie idle for three-quarters of the year, you’ve just jacked up the effective cost by 4X,” says Don Sadoway, an MIT chemist who cofounded Ambri, which has developed a liquid-metal grid battery that lasts about an hour longer than lithium-ion ones.

But it’s actually even worse. We’d need to overbuild renewables and storage to meet demand during the rarest events: the prolonged ebbs in sun or wind that happen every few years, maybe even once a decade.  

Regions don’t have to solve this problem entirely through storage. Meeting just a small share of total demand through other means would ease the cost targets that storage companies would need to reach, other research shows. That could include nuclear reactors, hydroelectric power, natural-gas plants with systems that capture carbon emissions, or long-distance transmission lines that can balance out renewables across time zones. But those options are politically unpopular, expensive, geographically constrained, or all three. Batteries have the advantage of not particularly bugging people.

We need to think about these future problems today because the necessary technologies could take years if not decades to develop. Areas with large shares of renewables, like California and Germany, already produce more solar or wind power than the grid can use during certain periods, undermining the economic incentives to build more. Many more regions are beginning to realize there’s a yawning gap that some technology will need to close if they hope to eliminate fossil fuels.

Developing cheap, long-duration batteries has stumped researchers for decades, mainly because the metals and chemicals that have worked best so far are expensive. Using them to meet longer storage needs means stacking up more and more of them. Form is guarded about its how it’s trying to sidestep these challenges, but part of the company’s approach is clear from a paper Chiang and colleagues published in the journal Joule in late 2017 (see “Serial battery entrepreneur’s new venture tackles clean energy’s biggest problem”).

All batteries contain two basic components: an electrolyte, usually a liquid chemical, and a pair of electrodes, the anode and the cathode, which are made of different materials (often, though not always, metals). Charged atoms, known as ions, carry current through the electrolyte between the two electrodes as the battery charges or discharges. In lithium-ion batteries, the electrolyte is some compound of lithium mixed with other chemicals.

In the 2017 paper, Chiang and his colleagues highlighted the potential of an “air-breathing aqueous sulfur flow battery.” A flow battery starts to get around the cost problem by separating the electricity-delivering components of the battery, including the electrodes, from the energy storage part, the electrolyte.

A standard flow battery has two different electrolytes, known as the catholyte and the anolyte, each of which can be stored in big, easily swapped tanks. So if you want more storage, you can just add larger tanks while those other pricey parts, including the electrodes, remain the same.

To make it really inexpensive, though, the electrolytes filling those giant tanks need to be cheap as well. The key to the flow battery in the Joule paper is to use a sulfur-based solution as the anolyte. Sulfur is among the most abundant elements in the earth’s crust as well as a by-product of fuel refining, so it’s extremely cheap and can store a lot of energy.

“Based on the charge stored per dollar, sulfur was more than a factor of 10 better than the next best thing,” Chiang told me in 2017.

Altogether, the chemical costs in such a flow battery could be as low as $1 per kilowatt-hour, according to the study.

When I spoke to Chiang last August, he confirmed that sulfur “is definitely still part of our road map.” He said it’s the approach they’re using in a project funded by the Department of Energy’s moonshot ARPA-E program. But Form says it’s now developing “multiple chemistries,” though it won’t say what the others are.

While most grid storage companies are focused just on the storage part, Jaramillo has also said they are exploring the possibility of “bidirectional power plants,” which would generate renewable energy on site using solar or wind, store it in big batteries, and deliver it to the grid as needed.

But an electrochemical battery, whether based on sulfur or lithium-ion chemistry or something else, is only one way of storing large quantities of energy.

In early September, a group of engineers crowded around a squat, silver cylinder about the size of a grill tank in the back of a cluttered workshop at Lawrence Berkeley National Lab, nestled in the hills looking over the San Francisco Bay. Aside from their intense gaze on the adjacent computer screen, the only hint that something was at work was an orange glow visible in a tiny window near the bottom of the device.

The researchers at Antora Energy are developing a new type of thermal storage. It’s a rarely used approach that retains energy in the form of extreme heat or cold in a variety of substances, like underground rocks or ice blocks. In Antora’s case, the substance inside the tank was a block of carbon that, at that moment, was running well above 2,000 ˚C.

The hope is they could use excess electricity from solar or wind farms to heat up that material, and then convert the heat back into electricity when it’s needed. Typically in thermal storage, this is still done in the highly inefficient 19th-century style: by creating steam that drives a turbine generator. But most of the energy gets wasted as a result of mechanical friction, steam leaks, and other issues.

Antora is testing a novel thermophotovoltaic system. It’s something like a solar panel, but it converts the infrared radiation coming off a hot object, rather than sunlight, into electricity. In late September, the researchers announced that they had set a new record by converting more than 30% percent of the heat flowing to the cell back into electricity in a lab experiment. They’re aiming to achieve more than 50% efficiency.

Mechanical methods offer another approach to grid storage. That includes pumping air into underground caverns, running rock-filled trains up hills, or transferring water between reservoirs at varying heights. All of these work in roughly the same way, using spare energy when it's available to move something to a higher elevation or place it under pressure. Then when it's released, we can harness the kinetic energy from the escaping air or descending trains or water to generate electricity.

Indeed, pumped hydro is by far our cheapest and most abundant source of grid energy storage today. The problem is you don’t always have enough water or hills near every power plant.

Under its “DAYS” program, ARPA-E has invested more than $30 million in 12 startups or research groups trying to crack the problem of grid storage. Those include Form’s flow batteries and Antora’s thermal system, as well as Quidnet Energy’s twist on pumped hydro: the San Francisco startup’s system pumps water into the gaps between confined rocks underground, creating pressure that forces the water back up and through a generator when electricity is needed.

Breakthrough Energy Ventures, the Bill Gates–backed fund, has made long-duration storage one of its highest priorities. In addition to Form, it has backed Quidnet and Malta, another thermal startup that relies on molten salt as the storage medium (see “Alphabet is in talks to spin out its molten-salt storage play”). 

Meanwhile, Japanese conglomerate SoftBank recently invested $110 million in the Swiss mechanical storage startup Energy Vault, which uses cranes and wires to stack up concrete blocks when renewables are generating excess electricity. It then drops those blocks back to the ground on those same wires, using their momentum to turn motors in the cranes in reverse and pump out electricity. (This video makes the concept clearer.)  

The unconventional nature of some of these ideas shows just how difficult a problem it is for technologies to make that leap from storing a few hours’ to a few weeks’ worth of energy.

“If we’re talking about capturing, say, one month or two months’ worth of energy during the summer and having it available for one month or two months in the winter, those are gigantic sums of energy,” Sadoway says. “How many train loads of rocks do you have?”

Most mechanical methods like trains or cranes require vast amounts of space. Thermal methods are inherently inefficient, since it’s hard to prevent the heat or cold from leaking away. And producing or burning most liquid fuels creates the very climate emissions we’re looking to avoid.

Batteries have the advantage of being clean, compact, mobile, and efficient. So if someone can make them cheap and long-lasting as well, they could plug into any grid. That’d enable wind and solar to provide far more of our electricity and, in turn, for clean electricity to meet much more of our total energy needs.

But those remain very big ifs. Some energy observers doubt Form can achieve its targets, or question how much natural gas such batteries would supplant even if they did. For their part, the company’s founders say it’s at least a decade-long project, with serious technical, financial, and market risks.

Major battery advances only occur about once every three decades on average, and the history of the field is cluttered with far more promising approaches that didn’t pan out than those that did.

Then again, the last one, lithium-ion technology, arrived on the market 28 years ago. We’re about due for another breakthrough.